Rong Huang

As the digitalization of power systems progresses, data-driven models have garnered widespread attention due to their performance advantages, leading to the emergence of numerous data-driven intelligent models for power tasks, such as attack detection and stability assessment. However, data-driven models are susceptible to adversarial attacks, even when deployed in highly secure control centers. Considering the similarity in the semantic features extracted by structurally diverse data-driven models when addressing the same downstream tasks, this paper proposes a transferable attention-distracting adversarial attack tailored for power systems. This attack first introduces an adversarial perturbation selection framework with physical constraints specific to power systems. It also offers different loss functions to distract attention and strategies to weaken the significance of features. Simulation experiments confirm that distracting the model’s attention results in more stable transferable attack effects and significantly reduces the performance of data-driven models across different task scenarios. The experimental results underscore the importance of not neglecting the security and robustness of models in security-critical scenarios like power systems, even while achieving optimal performance.

The network attack detection model based on machine learning (ML) has received extensive attention and research in PMU measurement data protection of power systems. However, well-trained ML-based detection models are vulnerable to adversarial attacks. By adding meticulously designed perturbations to the original data, the attacker can significantly decrease the accuracy and reliability of the model, causing the control center to receive unreliable PMU measurement data. This paper takes the network attack detection model in the power system as a case study to analyze the vulnerability of the ML-based detection model under adversarial attacks. And then, a mitigation strategy for adversarial attacks based on causal theory is proposed, which can enhance the robustness of the detection model under different adversarial attack scenarios. Unlike adversarial training, this mitigation strategy does not require adversarial samples to train models, saving computing resources. Furthermore, the strategy only needs a small amount of detection model information and can be migrated to various models. Simulation experiments on the IEEE node systems verify the threat of adversarial attacks against different ML-based detection models and the effectiveness of the proposed mitigation strategy.

Phasor measurement units (PMUs) used in the smart grids are vulnerable to time synchronization attacks (TSA), whose attack target is the global positioning systems (GPS). The attacker manipulates phase measurements by sending deceptive GPS signals to introduce errors in state estimation, which can affect bus power flow calculation and economic dispatching of the power system. Considering the case of multiple attacked PMUs in AC, a novel false phase data detection method is proposed in this paper. The method uses “phase coding” to extract and encode the potential relationship between amplitude and phase angle, and then builds a TSAs detection model based on a vector neural network that learns the encoded relationship vectors through dynamic routing. Finally, simulation experiments are carried out on IEEE 9, 14, 30, and 118 bus systems and real measurements provided by a province in China. The results show that the proposed method achieves the best performance among the detection methods based on the learning model.

The deployment and operation of load frequency control (LFC) systems rely on the latest network and communication technologies. This exposes it to network attacks, such as DoS attacks, which affect the running of the system. In order to offset the impact of DoS attacks, this article proposes a data prediction-based defense method. First, the cyber–physical model of LFC is constructed to derive the frequency. Based on the model, a prediction model organically combined with the hierarchical attention network is applied to restore missing measurements. Then the proposed method can detect and eliminate the impact of DoS attacks by comparing and replacing the measurements, rather than adopting flexible LFC strategies. Finally, simulation experiments are carried out on three different power generation systems. The results prove that the suggested method can effectively resist DoS attacks.

The rise of smart grid, which applies the Internet and smart metering technologies to the power system, results in massive, heterogeneous power data and a large number of attacks on the smart grid and causes great threats to the confidentiality and integrity of the power data. At the same time, with the development of quantum information, it is possible to use quantum cryptography to protect data transmission. Aiming at the problem of data security transmission in smart grid, this paper proposes a lightweight transmission scheme for power data security. In scheme, we combine quantum cryptography with " one-time pad " mechanism, use a quantum random number generator which compensates the defects of the traditional random number generator and quantum key distribution protocol which ensure the security of key distribution, and design a new lightweight stream cipher encryption algorithm. Through theoretical analysis and experiments, it is proved that the proposed transmission scheme can ensure the security of power data transmission.